Collaborative filtering recommender system in adversarial environment

Abstract

Collaborative filtering recommender system is wildly used in e-commerce system. According to the profiles of user or items, a collaborative filtering recommender system recommends items to targeted customers according to the preferences of their similar customers. It provides customer useful relevant information. Unfortunately, the recommender system is vulnerable to profile injection attacks. In the profile inject attack, the similar user profiles are manipulated by injecting a large number of fake profiles into the system. In this paper, four new attributes for the injection attack detection are proposed. We also discuss the profile injection attacks in adversarial learning environment. By applying the Localized Generalization Error Model (L-GEM), a more robustness attack profile detection system is proposed. Experimental results show that L-GEM based detection classifier has better robustness.